Personal Networks: Classic Readings and New Directions in Egocentric Analysis

Personal Networks, part of the Structural Analysis in the Social Sciences series, is an edited volume with a unique organizational scheme. Many edited volumes collect classic papers in a field and bind them together with minimal commentary. Such volumes are most useful when the interest is intellectual history of a field. Other edited volumes collect contemporary papers in a research space, often commissioned especially for the work, and offer a survey of topics currently attracting attention from researchers. These volumes are useful to orient beginners such as graduate students to the active research areas to which they might contribute. This volume does both and more. The editors Professors Small, Perry, Pescosolido, and Smith are to be commended!

After a very informative introduction, to be described shortly, the volume is divided into three sections: Early Foundations (seven selections – starting with Georg Simmel and ending with Harrison White), Later Foundations (twelve selections), and New Perspectives (nine contributions). Each Foundations selection, save one, is accompanied by a commentary that highlights the continuing relevance to contemporary work of the selection. Furthermore, the commentaries on the Later Foundations selections are often from the scholar who wrote the article, for example, Professor Ron Burt commenting on an excerpt from his influential 1992 book Structural Holes: The Social Structure of Competition. This innovation is refreshing, something that should be widely adopted for edited volumes of readings, and especially interesting when the commentary has the autobiographical twist.

The editors begin the introduction with the basic contrast between sociocentric and egocentric network research, then offer a brief history of ego-network research, and close with a section that characterizes the current era in egocentric research on five dimensions. As is likely well-known to readers of this review, sociocentric studies focus on complete networks of a population of actors where the ideal is that we know for every pair of actors if a tie exists between them or not. In the early days, this requirement meant sociocentric studies were limited to relatively small populations. Egocentric studies focus on (often) a sample of persons and their connections or particular types of ties to other individuals, where the ties these others have are unknown except in the extremely rare circumstance that they themselves are in the original sample. Egocentric studies can be thought of as an adaptation of sociocentric studies to the dominant data collection strategy in sociology—the survey of a random sample from a population. In the words of the editors (p. 4), “the power of sociocentric analysis comes from capturing the totality of social structure” while “the power of egocentric analysis comes capturing the entire set of people’s personal connections of any type.”

The editors divide the history of egocentric research into (a) “early notions” principally in the work of Simmel; (b) the work of survey researchers and mathematical sociologists, represented by Paul Lazarsfeld, James Coleman, Ed Laumann, White, Barry Wellman, and Claude Fischer; (c) contributions from ethnographers and anthropologists, represented in early work by Max Gluckman, John A. Barnes, Elizabeth Bott, and J. Clyde Mitchell and in later work by H. Russell Bernard and Peter Killworth; (d) research by psychiatrists and psychologists, Jacob Moreno, Fritz Heider, and Leon Festinger, Stanley Schacter, and Kurt Back; and, (e) more recent but still last century developments bringing in natural and physical scientists to what was (for them) a new field of “network science.” Not surprisingly, many of the key figures mentioned are authors of selections in the Early and Later Foundations sections.

Finally, five antipodal dimensions are used to contextualize current research: constraint vs agency, maps vs events, structure vs context, perception vs reality, and stasis vs dynamics. In each pairing, the first pole refers to a common tendency in past practice, while the second addresses an emerging concern in current practice. The constraint side of the first pairing emphasizes how network position shapes available resources and opportunities regardless of the intentions of occupants, while the agency side emphasizes how occupants may vary in their decision-making capability to access the resources and to take available opportunities. In the second pairing, maps refers to a presumption often made that a personal network consists of a stable set of connections and the first directive for the researcher is to map these links. Events refers to the idea that such a stable set of connections may not exist but that there is a field of possible links differentially activated by events, wherein the events that call up a response are as important as a mapping of the field of possibilities itself. The structure vs context theme evokes a comparison between approaches that abstract from the content of ties and the ontological nature of the vertices to focus on how the parts (vertices) are patterned via the edges (the ties) and approaches that place the content of ties and entity type of vertices as essential to the understanding of network mediated outcomes. The fourth antipodal dimension, perception vs reality, suggests a movement away from acceptance of self-report of connections at face value to a deeper understanding of the mechanisms by which actor store and retrieve information about connections to others. The last pairing, stasis vs dynamics, points out that past research faced practical limits—computational and otherwise—that necessitated a point-in-time snapshot of a network but that both computational and modelling resources have expanded tremendously, making over-time data collection and analysis feasible.

If this introductory material has a flaw, it is that it undersells the points being made by implying that they pertain to egocentric networks and not also to sociocentric network studies. To some extent this criticism applies to the rest of the collection as well. Much of the material is equally pertinent to sociocentric research.

Early Foundations includes excerpts from the writings of Simmel (commentary by Betina Hollstein), Elihu Katz and Lazarsfeld (commentary by Damon Centola), Mitchell (commentary by Borgatti and Halgin), Bott (commentary by Claire Bidart), Festinger, Schacter, and Back (commentary by John Levi Martin and Hyunku Kwon), Bernard, Killworth, David Kronenfeld, and Lee (commentary by Chris McCarty), and White (commentary by Ann Mische).

Hollstein provides a convincing argument that Simmel was not just a pure formalist, concerned with form or structural patterning of relationship, but was also concerned with the content of relations, theorized in terms of the “factors motivating individual action, such as drives, interests, purposes, and dispositions” (p. 49). She suspects the reputation as a pure formalist derives from the episodic way that English speaking sociological community was exposed to his ideas, noting that some of his work was only first translated in 1997. Her commentary goes a long way towards a more complete picture of Simmel’s theoretical framework, nicely represented in a two-page summary table (pp. 52f), that should be formatted and posted on reddit in r/coolguides.

Centola updates Katz and Lazarfeld with what we now know about social influence, particularly, the difference between simple information flows of rumor or ideas through a population, versus the complex diffusion of the adoption of a new behavior or convention. While Katz and Lazarfeld understood and popularized the role of social networks in information dissemination, recent work by Centola and colleagues makes equally clear that social networks are important to the diffusion of innovative behaviors or conventions, but just not in the same way. Information diffusion is a matter of simple contagion while behavior diffusion is a matter of complex contagion. In simple contagion, ordinary centrality matters and success depends on influencers being located in central cores; whereas in complex contagion, complex centrality matters and success depends on influencers being located in peripheries.

Borgatti and Halgin show how concepts and problems raised by Mitchell have their descendants in contemporary research. For instance, Mitchell identifies two generic problems that network thinking addresses, transmission of norms and goal achievement. Borgatti and Halgin find these problems reappear in current work into the foundations of social conformity, cohesion, and consensus, on one hand and on the other, into the understanding of social capital, resources embedded in network structure that make it possible to achieve objectives, both individual and collective, that could not be obtained without these resources. They also find “striking” Mitchell’s focus on norms in the article. Now, norms are the meat and potatoes of a cultural anthropologist, but Mitchell is after something more as Borgatti and Halgin see it: base relations (parenthood, marriage, friendship) are defined by normative expectations for attitudes, rights, duties, and behaviors and these expectations provide the context in which interactional events are understood as enactments of the base relations or idiosyncratic embroideries on them.

Bidart uses her research with colleagues, the Panel of Caen study, to reconsider Bott’s famous hypothesis about the degree of role segregation between husband and wife and the connectedness of the family’s network, specifically, the more connected, the more segregated, the more dispersed, the less segregated (more overlap). The Panel of Caen study is a qualitative panel (n=87) study of five waves from 1995 to 2015, collecting egocentric networks from a extensive series of name generators. Bidart shows that the general idea of Bott’s hypothesis, that role management in conjugal relations (and probably elsewhere) is associated with patterns of connections to other persons, couples, and groups holds up, but the exact details, including the micro-history of the relationship, matter and the overall picture is certainly more complex than originally envisioned. As Bidart (p. 132) concludes “Bott raised a beautiful question which, unless we oversimplify the parameters, remains, open, complex, and challenging.”

Martin and Kwon applaud Festinger, Schacter, and Back for having taken a step now widely recognized as essential to further progress in network analysis—namely, to conduct actual experiments with interventions to document clear causal connections from network ties to what actors know, believe, and act upon: they “carried out a combination of observational and real-world experiment that puts current work to shame” (p. 151). Martin and Kwon’s appraisal highlights the strengths of the work, namely, the focus on spatial proximity as a driver of social ties and analytical rigor in the conduct of a diffusion experiment to examine processes by which ideas and sentiments spread, but also its weaknesses, overoptimistic assumptions and a focus on social cohesion as a central concept. Much can be gained, they suggest, by retaining the strengths and building a more sophisticated conceptualization of social cohesion, allowing that group members may be held together by different types of forces (mechanically through homogeneity or organically through complementarity, to name two).

McCarty carefully summarizes the informant accuracy findings of Bernard, Killworth, Kronenfeld, and Sailer regarding informant reports on their social ties and the controversies elicited by the pithy conclusion that “about half of what informants report is probably incorrect in some way.” McCarty notes that some critics argued the conclusion followed from defective methodology; others sounded a more positive note, reanalyzing the datasets and finding agreement among respondents on the general pattern of social ties (being friends with) in the group (its social structure) than on the day-to-day occurrence of relational events on the level of behavior (eating lunch with). In either case, the informant accuracy studies and the studies challenging the findings raise dangers for personal network research that, McCarty suggests, can be mitigated by appropriate design of instruments: more accuracy is likely if instruments ask about salient and memorable types of social ties specific to context (who do you share needles with) rather than vague and ephemeral type of relational events (who did you talk to). Furthermore, he notes, well-designed instruments should (and must) remain the tool of choice, even in the age of social media scraping.

The final commentary in Early Foundations by Mische engages with excerpts from White’s Identity and Control from a unique point of view, as a former graduate student in a class taught by White titled “Identity and Control.” Her advice to readers of Identity and Control follows White’s own advice on dealing with challenging texts: parachute in and tame lions, i.e., pick a random start for exploration and make the author’s ideas work for your own questions. Her own foray into the wilds yields four summary ideas: “persons” are ephemeral contingent constructions; social life arises from intersecting struggles for stability and control; narratives beget networks and networks narratives; and ambiguity forces action in the service of contests for control. Direct implications for analysis of egocentric networks would recast them as standing waves of control attempts to manage identity formations.

Later Foundations includes excerpts from Fischer (commentary by Fischer), from Mark Granovetter (commentary by Roberto Fernandez), from Wellman and Scot Wortley (commentary by Wellman, Keith Hampton, Anabel Quan-Haase, and Molly-Gloria Harper), from Coleman (commentary by Beate Völker), from Pescosolido (commentary by Pescosolido), from Scott Feld (commentary by Feld, Devin T. Knighton and Alex McGail), from Burt (commentary by Burt), from Laumann, Peter V. Marsden, and David Prensky (commentary by Brea L. Perry and Adam R. Roth), from Miller McPherson, Lynn Smith-Lovin, and James M. Cook (commentary by McPherson, Smith-Lovin, and Craig Rawlings), from Robert Huckfeldt and John Sprague (commentary by Huckfeldt), from Nan Lin (commentary by Lin and Yanjie Bian), and, finally, a review written specially for this reader by Peter V. Marsden, Maleah Fekete, and Derick S. Baum.

In his commentary, Fischer provides a behind-the-scenes look at his premier egocentric network study, the 1977-1978 Northern California Community Study, highlighting its motivating concern with the impact of modernization on community. He contrasts it with the more recent 2015-2020 UC Berkeley Social Networks Project, another landmark egocentric study, improving on the earlier design but focused on a different problem, that of health outcomes. Fischer distills his vast experience with such substantial studies into a very useful and lengthy series of conceptual and methodological issues in ego-network research, modestly expressing the hope that “a few might inspire further attention” (p. 233).

“Strength of Weak Ties” has been cited so many times (55,525 as of 19 February 2020 according to Fernandez) that a commentary on it could go in many directions. Fernandez focuses on the labor market evidence Granovetter provided for the theory, namely, that weak ties were surprisingly useful for finding jobs. He surveys the state of labor market research on finding jobs, finding that causal evidence in favor of the theory is scarce and what evidence there is suggests the strength of weak ties lies in their abundance relative to strong ties. Causal evidence is scarce, Fernandez contends, for three reasons: selection on the dependent variable (information from those who found a job, rather than from all seekers, successful or not); changes in the dependent variable (the quality of job found rather than the fact of finding); and more sophisticated understanding of the complexity of the information chain linking job seeker, connecting agents, and hiring authority.

Wellman and colleagues in the commentary describe how the precarious status of “community” in the modern city was a driver of the egocentric East York studies they conducted with residents of a Toronto neighborhood. Four studies in all (1969, 1977-78, 2004-05, 2014), ranging from the pre-internet age to the age of smartphone ubiquity, document the communication habits and ties of social support enmeshing individuals in a web of community of family and friends. Wellman and colleagues note with some satisfaction that the findings of these studies put to rest pundits’ worries that urban life would destroy community while modern communication technology rendered it superfluous, concluding “although the structure and composition of community has changed, it has never been lost” (p. 292).

Völker’s commentary vividly describes how Coleman’s conceptualization of social capital has mushroomed into a variety of research programs. These programs try to identify just what makes a network beneficial to those embedded in it; to understand when social capital increases the returns, as Coleman argued, to human capital; to explore how social capital contributes to or ameliorates social inequality; and, most importantly, how social capital can be measured. Work in each of these areas must struggle with which aspects of social capital are at issue, identified by Völker by the cross-cutting axes of “access to” versus “use of” resources and actual provision versus hypothetical provision of those resources. Völker concludes that while much empirical research has demonstrated the value of Coleman’s concept, the overall theory is still underdeveloped, and many questions remain open for both sociocentric and egocentric studies.

Pescosolido’s commentary on her own 1992 article reminds us that developments in conceptualization and theory often result from the failure of attempts to frame very practical and personally relevant problems (health care seeking and decision making) with existing theoretical frameworks (rational choice). Pescosolido shows how her Social Organizational Strategy framework, an early attempt to embed the weighing of costs and benefits within a broader context of networks, blossomed into her current working model, the Network Embedded Symbiome, a multilevel conceptualization of social order that is networks all the way up to Place and Institutions and all the way down to Genes and Proteins. The implications of this conceptualization for egocentric studies are left to the reader’s imagination.

Commentary by Feld and colleagues on his foci theory emphasize the theory’s importance of structural mechanisms that produce regular patterns in networks of social ties, patterns such as disproportionate homophily or triadic closure. In their view, this lesson has yet to be fully appreciated in some quarters where tendencies exist to attribute such patterns to “unconstrained personal choice” (p. 368). They argue that advances in understanding network structural forms, including the favorite children of the new network scientists, “small worlds” and “scale free networks,” could profit much from systematic ways to identify relevant foci and the tie clusters they generate in a particular use case.

Burt’s commentary reminds us that often it is the absence of something that is important, in this case a tie between two of one’s associates, rather than its presence. His concept of “structural holes” is much broader than this simple example, of course, but the basic idea stands: the absence of ties gives us leverage on explaining some regularities that a focus on their presence would not. Burt’s commentary documents the many and varied empirical studies that have deployed the structural holes concept and its brokerage enhancing powers. Of special interest is the final section in which he maps out these concepts in terms of the five antipodal dimensions articulated by the editors in their introduction.

The boundary problem is a fundamental methodological problem for social research that network studies especially have been forced to contend with. The issue is: who belongs to the population whose ties are to be measured? Perry and Roth in their commentary point out that most of the citations (about 85%) to the foundational article on this topic by Marsden, Laumann, and Prensky come from sociocentric rather than egocentric studies. Their commentary draws out the implications of the article for egocentric studies. As they note, the boundary problem is both simpler and more complex for personal networks: on one hand, it is the views of the respondents that define the boundary of their social world, but on the other hand, there are many modes of connection and asking about them all, or even having an exhaustive list of the possibilities, is impossible so choices of name generators must be made. Their commentary illuminates the trade-offs that arise in practice when defining the foci of inclusion and provides some guidance on their resolution.

The commentary of McPherson, Smith-Lovin, and Rawlings explores just why the citations to the original homophily article continued to grow long past its publication. They suggest it occurred because, as the original article indicated, homophily – like associating with like over and beyond chance expectations – really is everywhere and so fits into a wide diversity of research programs in many fields, not just sociology. Perhaps their most important point, however, is the desirability of recapturing the sociological significance of homophily, namely, that “understanding homophily is really about understanding the forces of social cohesion and social fragmentation” (pp. 464f). Homophily gives us a lever for a deeper understanding of social structure, the principal aim of the sociological enterprise.

Political communication among citizens and its implications for the maintenance of democratic norms has never been more salient than it is today. Huckfeldt’s commentary on his work with Sprague first explains the background to the collection of egocentric networks of political discussion in the South Bend, IN study. But then he turns to what was learned from this and later studies that has contemporary currency for understanding the persistence of political disagreement and the resulting polarization of the body politic, identifying the crucial role of social networks: shared political identities require grounding in frequent and extensive social interaction channeled by our networks.

Lin and Bian’s commentary reviews Lin’s innovation in egocentric studies of social capital, the position generator probe. Respondents are asked if they know someone in various occupations that span the occupational hierarchy. The idea behind this innovation is that the resources available to an ego through their alters is a form of social capital equal to the structure of the network in which they are embedded (closed a la Coleman, or full of structural holes ala Burt). They suggest several directions for building on this idea: first, a focus on the reasons and motivations for why alters provide assistance as a bridging device from individual level to collective level social capital; and, second, developing a measure of organizational social capital similar to the position generator except the links to be reported are to organizations.

The final entry in the section “Later Foundations” is an original review for the volume that discusses the egocentric network items available in the General Social Survey. The most familiar of these items is the “discuss important matters with” item. Respondents provide up to five names and then answer questions about the named alters, the relationships to these alters, and a question about the alter-alter links. But there are other items that tap more aggregate aspects of a respondent’s ego networks, such as, items that ask for an estimate of the number of others known in a particular group or with a particular characteristic. The most valuable portion of this contribution is the review of the methodological research on the GSS egocentric items, research that bounds what we can and cannot infer with some confidence from the answers respondents give.

The final section New Perspectives includes original pieces on cognition, mobilization, trust, dynamics, inequality, culture, migration, movements, and social media. The first chapter in this section by Hui Sun, Matthew E. Brashears, and Edward B. Smith reviews the burgeoning research on how social networks are cognitively encoded by actors. This work intersects the informant accuracy work of an Early Foundations piece, seeking to understand the cognitive mechanisms in play when actors report on their ties to others. The big idea in the next chapter by Mario L. Small is the introduction of heterogeneity into analysis of what the literature calls “help seeking” or “social capital mobilization” or “tie activation.” Specifically, he addresses heterogeneity in terms of who instigates mobilization, ego or alter, whether the activation is planned or an accidental byproduct of everyday activity, and the source of assistance, another person or an organization. Small argues that understanding the how and when of these different mobilization events is important to understanding differences in wellbeing and social inequality. The third chapter by Sandra S. Smith and Jasmine M. Sanders also explores social capital mobilization but introduces the idea of self-verification as a motive for assistance, that is, one helps some egos but not others because those egos affirm in some way the self-views of the helper. The context for this exploration is job contacts assisting some job seekers and not others.

Brea L. Perry’s fourth chapter lays out six principles characteristic of complex systems, such as, “complex systems adapt to dynamic and uncertain environments” (p. 613), and then argues that personal networks fit all six and so can be viewed theoretically as complex systems. The payoff for this identification comes in Perry’s diagnosis of the methodological implications of personal networks being complex systems, implications that challenge existing practice, for example, extending interest beyond the core egocentric networks to include alters at the periphery and how the periphery interacts with the core. Han Shepherd and Filiz Garip in the fifth chapter introduce a model that exemplifies Perry’s complex systems approach. The model proposes a structure in which network characteristics, shaped by population features and organizational practices, differentially channel network processes, such as diffusion of resources and behaviors, to produce unequal outcomes for individuals and groups, for example, differences in academic achievement and migration decisions. A useful appendix to the chapter links the research literature to the steps in the model, identifying where various articles fit in the overall scheme.

The focus of Bonnie H. Erickson’s sixth chapter is to understand what cultural information is typically carried by weak ties. Weak ties are singled out because they are more likely to connect individuals in different social circles with divergent portfolios of cultural capital. The question Erickson poses is the circumstances under which such cultural capital flows through weak ties and the answer she proposes lays out three requirements that must be met for sharing to occur. These hypotheses are examined against data collected from ethnic status groups in Toronto, CN. The seventh chapter by Miranda Lubbers and José Luis Molina on migration notes the advances that have been made in understanding of the details of migrant’s networks using egocentric methods. These advances address three key levels of migration: how networks affect the volume and direction of migrant flows, how networks affect the migration decisions of individuals and their adaptation after moving, and how networks change in the process of assimilation in the host nation. The last question has an interesting twist to it as it points out how social capital is lost in the migration process and problematizes its regeneration in the migrant’s destination community.

Social movement research is a quintessential area to convince skeptics of the need for and the leverage provided by social network analysis because networks have long been recognized as key to all levels of movement activity: recruitment, support, communication, and political influence. In the eighth chapter, David Tindall, Mark C. J. Stoddart, John McLevey, Lorien Jasny, Dana R. Fisher, Jennifer Earl, and Mario Diani cover these aspects of social movements and the challenges their study presents for social network analysis: size in terms of large numbers of participants, lack of a sampling frame, secrecy about who is a social movement member, and serious issues of confidentiality when the movement is oppositional to the status quo. They review tools to address these problems and how qualitative and mixed methods research may be useful. Their last remarks on collective action as facilitated by virtual networks segue directly into the last chapter’s focus on social media.

It would be difficult to have a volume on personal networks that does not acknowledge the societal impact and the data revolution wrought by social media. The final chapter by Keith N. Hampton and Wenhong Chen discusses these issues in the context of egocentric research initiatives. Societal impact is captured by the concept of networked individualism and its affordances of persistent contact and pervasive awareness, that is, that participation in social media allows persons to remain in contact with others more easily and to have enhanced awareness of other’s activities even in the absence of communication exchange. The data revolution resides in the digital traces of activities that make it possible to have extraordinarily granular detail on who contacts whom or who is exposed to which topics of current interest. At the same time, the variety of social media platforms, the revision and updating of these platforms and their algorithms, and the differential participation of individuals in well and less well known platforms raise unique challenges for egocentric research.

Let me close this overlong review with a few general comments. First, the two sections on “Foundations” are likely to have multiple contributions of interest to anyone because (a) the topics covered are basic ideas in network analysis that have application in many substantive areas and (b) these ideas are important to researchers who study sociocentric networks as well as those who study egocentric networks. Second, in the “Perspectives” section, each article takes up a particular substantive area that may not be of concern to a wider audience. But if your research program examines culture or migration or social movements or any of the topics discussed, then you are well-advised to consult the relevant chapter. Third, if your research is sociocentric, do not ignore this volume just because the title appears to limit the material to egocentric analysis. Many if not most of the contributions are as applicable to sociocentric studies as to egocentric studies. Finally, this volume is about concept, theory, and intuition in the study of personal networks with key contributions on the methodology of data collection in the social network arena. In over 700 pages I saw only one reference to ERGMs. There are many other places to acquire the statistical expertise needed for the analysis of the complex interdependent objects presented by social networks. But if you are after motivating ideas and intuitions and theoretical frameworks—the theoretical side of the science of social networks—this volume is the place to start.